Natural Gas Load forecasting based on the fusion of VMD-SE and CNN-BiLSTM-Attention

Abstract The rational dispatch and steady supply of natural gas resources can be effectively supported by accurate forecasting the load. accuracy load is, however, low due to fundamental characteristics non-stationary non-linear Based on this, this paper proposes a new model based fusion VMD-SE CNN-BiLSTM-Attention. First, average instantaneous frequency method is used determine optimal K value VMD, then sample entropy introduced reconstruct sub-sequence VMD decomposition, which reduces computational overhead algorithm. Then, using CNN mine level information multi-dimensional time series variables, BiLSTM feature information, introducing Attention mechanism give corresponding weight different factors, highlighting influence key anticipated values each reconstructed sub-series are superimposed provide final predicted proposed in has lower error than other models, improve precision forecasting, better effect, according results establishing various models data one region China as an example.